Xailient’s Orchestrait platform is an AI lifecycle management tool for product companies working in the smart home space. This article explains what the Orchestrait platform is and discusses its many benefits.
The Orchestrait platform’s backstory
Xailient provides the world’s fastest Computer Vision (CV) Artificial Intelligence (AI) and specializes in putting accurate CV on tiny Edge AI devices.
Xailient’s technology is designed to make embedded Edge CV accurate, cost-effective, and real-time, and Orchestrait is designed to make using it (in production and at scale) easier and more insightful.
Orchestrait’s journey begins with Xailient’s patented Detectum algorithm.
Detectum is the world’s fastest and smallest object detector. It’s proven itself in the field many times, including our award-winning Purple Hive Project.
As Xailient became more efficient at managing our supercharged Detectum algorithm, we began building a suite of tools to help our internal commercial software engineering team support our various partners with their projects.
We understood that the suite of tools we created was paving the way to a more automated, observable, and repeatable method of empowering Edge CV at scale.
Most significantly, we realized that our new suite of tools provided a valuable alternative to what most of our partner companies were doing, which was manually managing their Edge CV operations.
Xailient wanted to provide an operational management toolkit that offered a holistic approach to managing Edge CV.
With our management toolkit, Xailient’s vision was to transform our algorithmic efficiency into a process that empowers companies to bring AI features and products to market easier, faster, and with less risk and resource investment. This process is a category of its own that’s called CVOps.
In continuing our work solving the most difficult, real-world CV problems for Edge AI devices, Xailient created the Orchestrait platform.
What exactly is the Orchestrait platform?
To put it simply, Orchestrait is a management tool for the smart home space. It allows product developers to better manage and monitor (in real-time) information about the Edge Face Recognition models running onboard all their deployed cameras.
Orchestrait is developing over-the-air updates for AI models and the deployment of add-on features.
Managing and monitoring algorithms in the field are essential in understanding how the algorithms respond to varying conditions and learning about application usage activities. Orchestrait monitors:
- Performance metrics (such as accuracy over time).
- Activity metrics (such as frequency of use over different periods, days, weeks, seasons).
- Customer feedback (sharing of accurate and inaccurate incidents).
This allows the Edge AI on your HomeCams and Smart Video Doorbells to evolve continuously and become smarter while avoiding manual monitoring.
When it comes to Orchestrait, speed of iteration is one of the things it does best.
Iteration refers to the process of repeating tasks within a design project. Iteration is a process that’s easier to carry out when there’s a standardized process and workflow to accelerate iteration cycles.
This is where Orchestrait shines, as quickly iterating on training data is key to building well-performing models for machine learning.
The many benefits of Xailient’s Orchestrait platform
Due to the global pandemic, Face Recognition cameras are now confronted with a new challenge, identifying people wearing face masks.
Thanks to face masks, customers report that their cameras can’t perform Face Recognition as accurately as before because the computer vision AI they’re using has been trained on only maskless faces.
By using the Orchestrait management platform, your camera can be updated to adapt to this situation so your customers can carry on enjoying your products even when changes like this occur.
This is why continual data training and in-field health and performance monitoring are so important for maintaining Edge AI and keeping it up to date. With the Orchestrait management platform, these processes are automated.
Moving from a manual process to automation offers big payoffs in speed, reliability, scalability, and quality assurance:
1. Speed
Orchestrait provides coordination and automation of all phases of the CV process. This drastically increases speed compared to the current manual processes used to manage and update Edge CV.
2. Reliability
The Orchestrait management platform is designed with fault tolerance and redundancy.
Fault tolerance refers to a system’s ability to continue to be dependable (both reliable and available) in the presence of inevitable component failures or subsystem failures. Redundancy is the duplication of important system functions.
Both components increase the platform’s reliability, making it well suited to mission-critical applications.
3. Scalability
Orchestrait is designed for scale to accommodate applications of all sizes. It can manage hundreds to millions of cameras.
4. Quality assurance
By streamlining the AI management process, the Orchestrait platform makes it consistent and repeatable. Additionally, checkpoints and monitors ensure quality KPIs are exceeded automatically.
In addition to the advantages listed above, Xailient’s Orchestrait also allows training data to be collected in an automated way that adheres to privacy best practices.
By redacting faces, Orchestrait enables the collection of privacy-compliant training data to build customer application training data sets.
This means your data collection will automatically comply with Europe’s General Data Protection Regulation (GDPR), California’s Consumer Privacy Act (CCPA), and other similar privacy regulations.
Orchestrait in action with a Face Recognition doorbell product
Let’s look at an example of how Orchestrait operates with a Face Recognition video doorbell product. The process would work like this:
1. Differentiate at the Edge
Deploy Edge Face Recognition with Orchestrait and upgrade your video doorbell products to identify visitors accurately.
2. Automate your data collection
Enable Orchestrait to start collecting video doorbell application training data in the field.
Now you always have an up-to-date, comprehensive front door training data set.
3. Innovate application AI
You can now use Orchestrait to train and innovate a brand-new detection feature, say delivery person detection at the front door, using your application training data set.
4. Deploy, monitor, rinse and repeat
Use Orchestrait to deploy your new delivery person detection to your video doorbell products and monitor its performance in the field. You can collect more training data and improve the algorithm. You own your training data set and the competitive differentiation you built on the platform.
Orchestrait delivers more value to your customers
Ultimately, Xailient’s Orchestrait platform allows you to reduce risk, rapidly deliver value, and save on costs. Orchestrait minimizes the time in bringing AI onto your cameras and accelerates your AI strategy so you can provide more value to your customers.
The Orchestrait platform makes AI development, management, and deployment easier, enabling more opportunities in new verticals.
About the Author
